Toxicology Letters 225 (2014) 401–406

Contents lists available at ScienceDirect

Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet

Mini review

Circulating extracellular vesicles as a potential source of new biomarkers of drug-induced liver injury Xi Yang, Zuquan Weng, Donna L. Mendrick, Qiang Shi ∗ Division of Systems Biology, National Center for Toxicological Research, Food and Drug Administration, 3900 NCTR Road, Jefferson, AR 72079, USA

h i g h l i g h t s • Hepatocytes produce and release extracellular vesicles to circulation. • Circulating extracellular vesicles contain molecules reflecting drug hepatotoxicity. • Hepatotoxicity biomarkers from extracellular vesicles need to be validated.

a r t i c l e

i n f o

Article history: Received 9 December 2013 Received in revised form 7 January 2014 Accepted 8 January 2014 Available online 21 January 2014 Keyword: Drug-induced liver injury Biomarkers Extracellular vesicles Exosomes

a b s t r a c t Like most cell types, hepatocytes constantly produce extracellular vesicles (EVs) such as exosomes and microvesicles that are released into the circulation to transport signaling molecules and cellular waste. Circulating EVs are being vigorously explored as biomarkers of diseases and toxicities, including druginduced liver injury (DILI). Emerging data suggest that (a) blood-borne EVs contain liver-specific mRNAs and microRNAs (miRNAs), (b) the levels can be remarkably elevated in response to DILI, and (c) the increases correlate well with classical measures of liver damage. The expression profile of mRNAs in EVs and the compartmentalization of miRNAs within EVs or other blood fractions were found to be indicative of the offending drug involved in DILI, thus providing more informative assessment of liver injury than using alanine aminotransferase alone. EVs in the urine and cell culture medium were also found to contain proteins or mRNAs that were indicative of DILI. However, major improvements in EV isolation methods are needed for the discovery, evaluation, and quantification of possible DILI biomarkers in circulating EVs. Published by Elsevier Ireland Ltd.

Contents 1. 2. 3. 4.

5.

6. 7. 8.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The constituents of EVs from hepatocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Circulating EVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Blood EVs are a promising source for DILI biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. EV mRNA as biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. EV miRNA as biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. EV proteins as possible biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Urinary EVs are a new source for DILI biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. EV miRNA as biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. EV proteins as biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EVs in cell culture medium are a useful source of DILI biomarkers for in vitro models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Advantages and disadvantages in using EVs as DILI biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Disclaimer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

∗ Corresponding author. Tel.: +1 870 543 7365; fax: +1 870 543 7736. E-mail address: [email protected] (Q. Shi). 0378-4274/$ – see front matter. Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.toxlet.2014.01.013

402 402 402 403 403 403 403 403 403 403 403 404 405 405 405 405

402

X. Yang et al. / Toxicology Letters 225 (2014) 401–406

Fig. 1. Secreted extracellular vesicles (EVs). Both exosomes and microvesicles are actively released from cells. (1) Exosomes, originating from multivesicular endosomes (MVEs), are secreted as a consequence of fusion with plasma membrane. Some MVEs fuse with the lysosome. (2) Microvesicles are directly derived from plasma membrane. The constituents of EVs include nucleic acids (comb symbols), proteins (purple), and lipids (red).

1. Introduction Diagnostic, therapeutic, and prognostic tools for severe druginduced liver injury (DILI) are limited and imperfect (Hussaini and Farrington, 2014; Regev, 2013), and thus it is not surprising that efforts to identify new biomarkers of DILI continue to grow. A new (or novel) DILI biomarker herein refers to those that are different from the conventionally used four serum biomarkers (alanine aminotransferase (ALT), aspartate aminotransferase, alkaline phosphatase, and bilirubin) that are recently endorsed by the FDA for pre-market clinical safety evaluation (FDA, 2009). The major drawbacks of the conventional DILI biomarkers include low organ specificity and poor reflection of liver functions (FDA, 2009). In recent years, the introduction of novel experimental designs such as detecting pre-exposure baseline gene or protein profiles (Borlak et al., 2013; Yun et al., 2010) and new molecular entities such as microRNAs (miRNAs) (Starkey Lewis et al., 2011; Wang et al., 2009) and mechanistic biomarkers (Antoine et al., 2012, 2013; in combination with cutting-edge technologies such as systems biology-based approaches) have significantly advanced the discovery of new DILI biomarkers. However, their possible acceptance in clinical and regulatory settings still needs additional time and effort. Since these advances have been reviewed elsewhere (Aubrecht et al., 2013; McGill and Jaeschke, 2013; Mendrick, 2011; Shi et al., 2010, 2013; Starkey Lewis et al., 2011), no details will be provided here. A poorly explored but seemingly promising approach is to use extracellular vesicles (EVs) to discover new DILI biomarkers. Here is a brief review on recent advances and potential issues in this scientific endeavor. To our knowledge, this is the first review about discovering circulating EVs-based DILI biomarkers. For a general overview of EVs in liver physiology and pathophysiology, the readers are referred to two recently published reviews (Masyuk et al., 2013; Royo and Falcon-Perez, 2012). EVs are generally referred to as membrane-surrounded structures, constantly released by almost all types of cells, and have been classified into several sub-categories. Examples include exosomes (50–100 nm in diameter) that originated from endosomal membranes and microvesicles (20–1,000 nm in diameter) that are derived from plasma membranes (van der Pol et al., 2012). A schematic diagram of EVs is presented in Fig. 1. Apoptotic bodies (50–5,000 nm in diameter) that are formed only during programmed cell death are also a sub-type of EVs. However, a

consensus nomenclature for all sub-types of EVs is not currently available (van der Pol et al., 2012). It is likely that most of the reported sub-types of EVs were actually a heterogeneous mixture of different origins, sizes, and constituents. It is for this reason that the term EVs is used in the present review, instead of specific terms such as exosomes that were used in some original references. The list of known functions of EVs continues to expand. The consensus seems to be that EVs are used by essentially all organisms to transport and deliver various substances including signaling molecules and cellular waste (Raposo and Stoorvogel, 2013; van der Pol et al., 2012). Exogenously inserted molecules such as drugs may be transported by EVs without compromising their effectiveness (Raposo and Stoorvogel, 2013; van der Pol et al., 2012). Though the mechanism remains unclear, the structure and key functions of proteins and nuclear acids in circulating EVs are generally wellpreserved (Raposo and Stoorvogel, 2013; van der Pol et al., 2012). This unique character makes EVs particularly suitable for developing non-invasive or minimally invasive biomarkers, such as those from the urine or from the blood, respectively. The constituents of EVs include proteins, nuclear acids, and lipids. When cells undergo stress, the EV constituent changes accordingly (de Jong et al., 2012; Kucharzewska et al., 2013) and forms the basis of using EVs as disease and toxicity biomarkers. The well-established DILI biomarker ALT has a half time of about 13 h in the circulation (Ribeiro et al., 2003), and the newly proposed DILI biomarker miR-122 has an even shorter half time of several hours (Starkey Lewis et al., 2011; Wang et al., 2009). False negatives are possible when a critical window of injury detection is missed by using short-lived biomarkers. Circulating EVs-based DILI biomarkers are expected to minimize or avoid this issue, as macromolecules in EVs are protected from degradation. Nevertheless, biomarkers with a long half-life may not be informative as to injury repair. 2. The constituents of EVs from hepatocytes Deciphering the constituents of hepatocyte-derived EVs helped pave the way for discovering EV-based DILI biomarkers. The proteins and mRNA species in hepatocyte EVs have been explored using systems biology-based approaches. In EVs isolated from primary rat hepatocytes, up to 251 proteins were identified, among which at least 109 (43%) were identical to those found in EVs of non-hepatocyte origins (Conde-Vancells et al., 2008). Some of the other proteins may be specific to hepatocyte-derived EVs, but additional work is needed to confirm it. Gene microarray analysis showed that EVs from rat hepatocytes contained 1336 gene transcripts (Royo et al., 2013). Pathway analysis found that the proteins and genes identified in hepatocyte EVs were enriched in xenobiotic metabolism and energy production, as well as other processes mediated by the liver (Conde-Vancells et al., 2008; Royo et al., 2013). Unfortunately, though the lipid composition of EVs of nonhepatocyte origin has been examined using a lipidomics approach (Llorente et al., 2013), such data are not available for hepatocytederived EVs. Similarly, a systems biology analysis of miRNAs in EVs from hepatocytes is lacking. 3. Circulating EVs EVs have been found in various types of body fluids, such as blood (serum/plasma), urine, saliva (Lasser et al., 2011a), cerebrospinal fluid (Street et al., 2012), bronchoalveolar lavage fluid (Levanen et al., 2013), nasal secretions (Lasser et al., 2011b), breast milk (Lasser et al., 2011a), semen (Poliakov et al., 2009), etc. Though the exact tissue origin of circulating EVs is difficult to ascertain, it is generally believed that they are a mixture of EVs from many

X. Yang et al. / Toxicology Letters 225 (2014) 401–406

different cell types (de Vooght et al., 2013; Huang et al., 2013). Of note, it was found that EVs of normal rat blood contain liver-specific mRNAs, indicating that hepatocyte-derived EVs are released into the circulation under physiological conditions (Wetmore et al., 2010). 4. Blood EVs are a promising source for DILI biomarkers 4.1. EV mRNA as biomarkers As is true with ALT, liver-specific mRNAs are released from damaged liver tissue to the circulatory system and therefore useful to assess liver injury (Miyamoto et al., 2008; Okubo et al., 2013). However, one possible drawback is that many released mRNAs are rapidly degraded in the blood if they are in the naked form. This issue could be at least partially resolved by detecting mRNAs wrapped in EVs. The first study in this direction was published in 2010 using a rat model of d-(+)-galactosamine-induced acute liver injury (Dg-ALI) (Wetmore et al., 2010). The expression level of some liver-specific mRNAs (e.g., albumin (Alb), fibrinogen B ˇ-polypeptide (Fgb), haptoglobin (Hp), and ˇ-actin (Actb)) was significantly increased in plasma EV fractions and generally reduced in liver tissue, suggesting that mRNAs of liver origin were released into blood and packed into plasma EVs in response to liver damage (Wetmore et al., 2010). Such increases showed a good correlation to elevations of ALT and the histopathological assessment of liver injury (Wetmore et al., 2010). Interestingly, a hepatotoxic challenge by d-(+)-galactosamine caused not only a change in the size of plasma EVs, but also the mRNA distribution pattern in different types of EVs (Wetmore et al., 2010). Given the important functions of EVs in cell-to-cell communication and in mediating immune responses, further investigations on these changes in plasma EVs are likely to provide novel mechanistic insights into DILI pathogenesis as well. Of note, these liver-specific mRNAs were also readily detectable in EVs from normal, untreated rats (Wetmore et al., 2010). This study provided a proof-of-concept that a change in liver physiology and onset of liver damage would trigger perturbations in blood EVs, and therefore more liver-specific DILI biomarkers could be discovered in circulating EVs. The above-mentioned finding was reproduced and expanded in a more recent report. Using a Dg-ALI model, it was found that serum EVs from treated rats contained significantly increased Alb mRNAs as reported previously, as well as three other mRNAs, guanine nucleotide binding protein (G protein), beta polypeptide 2-like 1 (Gnb2l) and retinol binding protein 4, plasma (Rbp4) (Royo et al., 2013). Possibly due to the differences in EV isolation protocol, this study did not find Alb in the serum of normal, untreated rats (Royo et al., 2013) while the previous study did see its presence in plasmaderived EVs (Wetmore et al., 2010).

403

miR-122 and miR-155 were predominantly detected in extra-EV fractions but not the EV fractions(Bala et al., 2012). This is the first study suggesting that miRNA distribution patterns inside or outside of EVs may reflect DILI etiology. Interestingly, the miRNA expression pattern in human blood EVs was also found to be changed in patients with chronic hepatitis-C or non-alcoholic steatohepatitis, and such changes were reflective of the severity of liver damage (Murakami et al., 2012). As DILI could mimic all types of liver diseases (FDA, 2009), these newly identified biomarkers of liver disease may also be used as novel DILI biomarkers. 4.3. EV proteins as possible biomarkers Proteins from blood EVs have not been examined in DILI model. However, in patients with chronic hepatitis-C, the protein soluble CD81 was remarkably increased in serum EVs and the elevation correlated with ALT (Welker et al., 2012), suggesting that protein changes in blood EVs may also occur in response to DILI. 5. Urinary EVs are a new source for DILI biomarkers 5.1. EV miRNA as biomarkers Though urine is not an intuitive source of DILI biomarkers, work done here recently found that a subset of urinary miRNAs was increased in a rat model of acetaminophen or carbon tetrachloride-induced liver injury (Yang et al., 2012). The unique advantage of using urinary DILI biomarkers is that urine collection is non-invasive, samples can be obtained in large quantities and subjects could be easily assayed in time-course studies. As the isolation methods of urinary EVs are being vigorously defined and refined, and urinary EVs have been shown to carry different types of biomarkers for kidney diseases (Alvarez et al., 2012; Gonzales et al., 2009), it appears highly plausible to use urinary EVs for discovering new DILI biomarkers. 5.2. EV proteins as biomarkers A recent, first of its kind, study found in a rat model of DgALI that several proteins in urinary EVs, such as Cd26 (DPP4), Slc3a1 (rBAT), Cd81, lysosome membrane protein II (LimpII), and Cd10, were remarkably reduced in rats with histopathologically confirmed liver damage (Conde-Vancells et al., 2010). Though a performance comparison was not made between traditional blood DILI biomarkers and urinary EV-based protein alterations in detecting liver damage, this study clearly suggests that urinary EVs are a new source for discovering novel DILI biomarkers. 6. EVs in cell culture medium are a useful source of DILI biomarkers for in vitro models

4.2. EV miRNA as biomarkers In addition to liver-specific mRNAs, some liver-enriched miRNAs such as miR-122 were also found to be good candidates of new DILI biomarkers, as has been recently reviewed (Shi et al., 2013). In circulating EVs, two liver-enriched miRNAs have been suggested to be more informative DILI biomarkers than ALT alone (Bala et al., 2012). Specifically, in mouse models of alcohol-induced liver injury and lipopolysaccharide challenge-induced inflammatory liver injury, serum and plasma levels of miR-122 and miR-155 were remarkably increased and almost exclusively detected in circulating EVs but not in the non-EV fractions. However, a reverse distribution pattern was observed in the acetaminophen-induced liver injury (AILI) model; that is, elevated circulating levels of

Primary cultured hepatocytes are a useful in vitro DILI model (Godoy et al., 2013; Shi et al., 2011) and are widely used in the drug development process to study drug metabolism (FDA, 1997). It was recently found that several mRNAs in hepatocyte-derived EVs were remarkably increased in response to treatment by hepatotoxic drugs such as acetaminophen, diclofenac, and d-(+)-galactosamine, and the changes appeared to be drug-specific (Royo et al., 2013), which is reminiscent of drug-specific changes in EV miRNA under in vivo conditions (Bala et al., 2012). Of note, acetaminophen did not significantly affect mRNAs in EVs from cell culture medium (Royo et al., 2013), just as it did not remarkably increase miR-122 in serum EVs (Bala et al., 2012), indicating that EVs might not be a good tool to detect AILI.

404

X. Yang et al. / Toxicology Letters 225 (2014) 401–406

Table 1 Summary of reported EV-based DILI biomarkers. Drug/chemical tested

Model used

Sample type

EV isolation method

Proposed biomarkers

References

d-(+)-galactosamine

Rats/in vivo

Plasma

Ultracentrifugation

Wetmore et al. (2010)

d-(+)-galactosamine, LPS, and acetaminophen d-(+)-galactosamine

Mice/in vivo

Plasma and serum

ExoQuick kit

Liver-specific mRNAs: Alb, Fgb, Hp, and Actb Liver specific miRNAs: miR-122, miR-155

Rats/in vivo

Serum

ExoQuick kit

Royo et al. (2013)

d-(+)-galactosamine

Rats/in vivo

Urine

Ultracentrifugation

d-(+)-galactosamine, diclofenac, and acetaminophen

Rats/in vitro (primary hepatocytes)

Cell culture medium

ExoQuick kit

mRNAs: Alb, Gnb2l and Rbp4 Proteins: Cd26 (DPP4), Slc3a1 (rBAT), Cd81, LimpII and Cd10 mRNAs: Alb, Gnb2l and Rbp4

Bala et al. (2012)

Conde-Vancells et al. (2010)

Royo et al. (2013)

Alb, albumin; Fgb, fibrinogen B ␤-polypeptide; Hp, haptoglobin; Actb, ␤-actin; Gnb2l, guanine nucleotide binding protein (G protein), beta polypeptide 2-like 1; Rbp4, retinol binding protein 4, plasma; LPS, lipopolysaccharide; LimpII, lysosome membrane protein II.

Although in vitro models generally have low predictability of human DILI, the recent introduction of combined measurements of multiple endpoints appears to be able to significantly increase the correlation between in vitro data and a drug’s potential to cause human hepatotoxicity (Tolosa et al., 2012; Xu et al., 2008). It appears reasonable to speculate that the incorporation of EV-based biomarkers may further enhance the power of in vitro models in predicting human DILI. Manipulation of cells in culture is far easier than animal or human experiments. Thus, using this scenario in a discovery mode might provide good candidates to be examined under in vivo conditions. Indeed, some of the mRNAs identified in vitro were also confirmed in in vivo models (Royo et al., 2013). All the above-mentioned EV-based DILI biomarkers are summarized in Table 1. An overview of the strategies for EV isolation and downstream analysis was presented in Fig. 2. 7. Advantages and disadvantages in using EVs as DILI biomarkers There are several advantages in using EV-based DILI biomarkers. First, it may be possible to identify highly specific DILI biomarkers in circulating EVs. This is important as currently used DILI biomarkers lack organ specificity (FDA, 2009). Second, EV components are protected from degradation (Raposo and Stoorvogel, 2013; van der Pol et al., 2012) and thus have a longer survival time, providing a wider time window for injury detection. Third, efforts to directly

Fig. 2. An overview of EV isolation and downstream analysis for discovering novel DILI biomarkers. PBS, phosphate buffered saline; Q-PCR, quantitative polymerase chain reaction.

use serum and plasma to discover new DILI biomarkers are often hindered by highly abundant blood constituents. The true signal may escape detection among those very strong and irrelevant background signals. Using isolated EVs may avoid this problem. While the above advantages appear generally applicable, one additional benefit of using EVs in DILI is that not only changes in their sizes, but also patterns of their association with specific molecules may be indicative of the etiology of liver injury (Bala et al., 2012). Though the concept of gene expression patterns in EVs being specific to the toxicant or injury type is not necessarily new, the added dimension of comparing expressions of mRNAs and miRNAs inside or outside EVs may provide further insights into the mechanisms of DILI and additional evidence in ascertaining the identity of the drug causing DILI (Bala et al., 2012; Wetmore et al., 2010). Disadvantages in using EV-based DILI biomarkers are also very apparent. The most important one is that the EV isolation method has not been standardized (Momen-Heravi et al., 2013). In fact, sub-populations of EVs are not even well defined. The current “gold standard” isolation method is based on ultracentrifugation (Momen-Heravi et al., 2013) which requires costly instrumentation, at least 2 h to complete the isolation, and large amounts of blood or urine. This obviously is not suitable in clinical settings. New methods have been developed to overcome these problems, but a vigorous evaluation has not been performed. Usually these newly developed methods yield only a small amount of EVs, making it not possible to fully characterize the end product. Some isolation methods also tend to enrich a specific type of molecules and this has been claimed to be a unique advantage of several commercial products (Channavajjhala et al., 2013; Eldh et al., 2012). Though this might be advantageous for certain specific applications, the importance of charactering the isolated “EVs” using traditional methods such as electron microscopy shall not be underestimated. After EVs are isolated, methods to further extract nuclear acids also need to be optimized, as different extraction methods have been shown to significantly affect the yield and patterns of EV mRNAs and miRNA (Channavajjhala et al., 2013; Eldh et al., 2012). The same is also true for proteomics analysis of EVs (Kalra et al., 2013). At this stage, the best approach one can adopt is to use a consistent isolation procedure across all test samples in a specific lab. However, the matter of inter-laboratory validation remains to be resolved and thus limits the usefulness of any discovered biomarkers. In addition to the isolation method, data normalization methods also need improvement. Most studies did not even mention this issue, likely because it was assumed that an identical starting sample for EV isolation provided an adequate control. A “housekeeping” molecule whose expression level is consistent and resistant to pathophysiological changes is highly desirable but not currently available for EVs. The last disadvantage is that the dynamic change of EVs, especially

X. Yang et al. / Toxicology Letters 225 (2014) 401–406 Table 2 Advantages and disadvantages of using circulating EV-based DILI biomarkers. Advantages in using EVs as DILI biomarkers

Disadvantages in using EVs as DILI biomarkers

1. High sensitivity due to exclusion of abundant molecules and enrichment of specific components 2. High liver specificity

1. Relatively large sample volume needed (not suitable for small animals in pre-clinical study)

3. May reflect liver etiology 4. Candidate biomarkers are protected from degradation

2. Difficult and time-consuming to isolate EVs (usually needs ultracentrifuge and takes >2 h) 3. Dynamics of EVs not well-understood 4. Long half-lives suggest it may be difficult to ascertain if and when the injury is reversing 5. Data normalization method poorly studied

those released into circulation after tissue damage, is still poorly understood (Raposo and Stoorvogel, 2013; van der Pol et al., 2012). The advantages and disadvantages in using EVs as DILI biomarkers are summarized in Table 2. 8. Conclusions Discovering EV-based DILI biomarkers is just in its infancy. For circulating EV-based biomarkers to be of practical use, an easy to perform method of EV isolation is the first essential step. Efforts are ongoing to standardize the isolation method of EVs (Witwer et al., 2013). Hopefully a consensus method will be established in the next few years. As shown in Table 1, all publications to date are based on animal studies. To ascertain the usefulness of circulating EVs in human DILI is undoubtedly an interesting future direction. Though a mice study showed EV miRNAs were indicative of the etiology of liver injury (Bala et al., 2012), whether this is true in human needs careful evaluation. This is particularly important since the distribution pattern of miRNA in human blood EVs is highly controversial, with one study showing that miRNA in human plasma locate mainly outside of EVs (Turchinovich et al., 2011) and other investigations showing the contrary (Arroyo et al., 2011; Gallo et al., 2012). In addition, only a few drugs had been investigated (Table 1). To further qualify such biomarkers, ascertain if they are drug-specific and/or distinguish mechanisms of injury, it will be essential to study a large number of drugs in the future studies. DILI could be broadly classified into three types, that is, hepatocellular, cholestatic, and mixed types (FDA, 2009; Shi et al., 2010). Though the perturbation pattern in rat circulating miRNAs has been shown to be reflective of these different liver injury types (Yamaura et al., 2012), it is unknown if EVs have the same potential. DILI could also be arguably categorized into “intrinsic” and “idiosyncratic” (Corsini et al., 2012). Acetaminophen is generally believed to be an “intrinsic” hepatotoxicant whose pathogenesis can be readily mimicked by animal models. However, for drugs that cause “idiosyncratic” liver injury, animal models are usually not available (Corsini et al., 2012). It is likely for this reason that efforts to discover EV-based biomarkers are currently limited to drugs that cause “intrinsic” DILI. It remains to be investigated if EVs are useful in detecting “idiosyncratic” DILI in the future. Disclaimer This article is not an official guidance or policy statement of the US Food and Drug Administration (FDA). No official support or endorsement by the FDA is intended or should be inferred.

405

Acknowledgements This study is supported by the US FDA’s Chief Scientist’s Challenge Grants program. Dr. Zuquan Weng is supported by the Research Participation Program at the National Center for Toxicological Research administrated by the Oak Ridge Institute for Science and Education through an interagency agreement between the US Department of Energy and the US Food and Drug Administration. References Alvarez, M.L., Khosroheidari, M., Kanchi Ravi, R., DiStefano, J.K., 2012. Comparison of protein, microRNA, and mRNA yields using different methods of urinary exosome isolation for the discovery of kidney disease biomarkers. Kidney Int. 82, 1024–1032. Antoine, D.J., Jenkins, R.E., Dear, J.W., Williams, D.P., McGill, M.R., Sharpe, M.R., Craig, D.G., Simpson, K.J., Jaeschke, H., Park, B.K., 2012. Molecular forms of HMGB1 and keratin-18 as mechanistic biomarkers for mode of cell death and prognosis during clinical acetaminophen hepatotoxicity. J. Hepatol. 56, 1070–1079. Antoine, D.J., Dear, J.W., Lewis, P.S., Platt, V., Coyle, J., Masson, M., Thanacoody, R.H., Gray, A.J., Webb, D.J., Moggs, J.G., Bateman, D.N., Goldring, C.E., Park, B.K., 2013. Mechanistic biomarkers provide early and sensitive detection of acetaminophen-induced acute liver injury at first presentation to hospital. Hepatology 58, 777–787. Arroyo, J.D., Chevillet, J.R., Kroh, E.M., Ruf, I.K., Pritchard, C.C., Gibson, D.F., Mitchell, P.S., Bennett, C.F., Pogosova-Agadjanyan, E.L., Stirewalt, D.L., Tait, J.F., Tewari, M., 2011. Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma. P. Natl. Acad. Sci. USA 108, 5003–5008. Aubrecht, J., Schomaker, S.J., Amacher, D.E., 2013. Emerging hepatotoxicity biomarkers and their potential to improve understanding and management of drug-induced liver injury. Genome Med. 5, 85. Bala, S., Petrasek, J., Mundkur, S., Catalano, D., Levin, I., Ward, J., Alao, H., Kodys, K., Szabo, G., 2012. Circulating microRNAs in exosomes indicate hepatocyte injury and inflammation in alcoholic, drug-induced, and inflammatory liver diseases. Hepatology 56, 1946–1957. Borlak, J., Chatterji, B., Londhe, K.B., Watkins, P.B., 2013. Serum acute phase reactants hallmark healthy individuals at risk for acetaminophen-induced liver injury. Genome Med. 5, 86. Channavajjhala, S.K., Rossato, M., Morandini, F., Castagna, A., Pizzolo, F., Bazzoni, F., Olivieri, O., 2013. Optimizing the purification and analysis of miRNAs from urinary exosomes. Clinical chemistry and laboratory medicine: CCLM/FESCC. 52, 1–10. Conde-Vancells, J., Rodriguez-Suarez, E., Embade, N., Gil, D., Matthiesen, R., Valle, M., Elortza, F., Lu, S.C., Mato, J.M., Falcon-Perez, J.M., 2008. Characterization and comprehensive proteome profiling of exosomes secreted by hepatocytes. J. Proteome Res. 7, 5157–5166. Conde-Vancells, J., Rodriguez-Suarez, E., Gonzalez, E., Berisa, A., Gil, D., Embade, N., Valle, M., Luka, Z., Elortza, F., Wagner, C., Lu, S.C., Mato, J.M., Falcon-Perez, M., 2010. Candidate biomarkers in exosome-like vesicles purified from rat and mouse urine samples. Proteomics Clin. Appl. 4, 416–425. Corsini, A., Ganey, P., Ju, C., Kaplowitz, N., Pessayre, D., Roth, R., Watkins, P.B., Albassam, M., Liu, B., Stancic, S., Suter, L., Bortolini, M., 2012. Current challenges and controversies in drug-induced liver injury. Drug Saf. 35, 1099–1117. de Jong, O.G., Verhaar, M.C., Chen, Y., Vader, P., Gremmels, H., Posthuma, G., Schiffelers, R.M., Gucek, M., van Balkom, B.W., 2012. Cellular stress conditions are reflected in the protein and RNA content of endothelial cell-derived exosomes. J Extracell Vesicles. 1. de Vooght, K.M., Lau, C., de Laat, P.P., van Wijk, R., van Solinge, W.W., Schiffelers, R.M., 2013. Extracellular vesicles in the circulation: are erythrocyte microvesicles a confounder in the plasma haemoglobin assay? Biochem. Soc. Trans. 41, 288–292. Eldh, M., Lotvall, J., Malmhall, C., Ekstrom, K., 2012. Importance of RNA isolation methods for analysis of exosomal RNA: evaluation of different methods. Mol. Immunol. 50, 278–286. FDA, 1997. Guidance for Industry: Drug Metabolism/Drug Interaction Studies in the Drug Development Process: Studies In Vitro. FDA, 2009. Guidance for Industry: Drug-Induced Liver Injury: Premarketing Clinical Evaluation. Gallo, A., Tandon, M., Alevizos, I., Illei, G.G., 2012. The majority of microRNAs detectable in serum and saliva is concentrated in exosomes. PloS One 7, e30679. Godoy, P., Hewitt, N.J., Albrecht, U., Andersen, M.E., Ansari, N., Bhattacharya, S., Bode, J.G., Bolleyn, J., Borner, C., Bottger, J., Braeuning, A., Budinsky, R.A., Burkhardt, B., Cameron, N.R., Camussi, G., Cho, C.S., Choi, Y.J., Craig Rowlands, J., Dahmen, U., Damm, G., Dirsch, O., Donato, M.T., Dong, J., Dooley, S., Drasdo, D., Eakins, R., Ferreira, K.S., Fonsato, V., Fraczek, J., Gebhardt, R., Gibson, A., Glanemann, M., Goldring, C.E., Gomez-Lechon, M.J., Groothuis, G.M., Gustavsson, L., Guyot, C., Hallifax, D., Hammad, S., Hayward, A., Haussinger, D., Hellerbrand, C., Hewitt, P., Hoehme, S., Holzhutter, H.G., Houston, J.B., Hrach, J., Ito, K., Jaeschke, H., Keitel, V., Kelm, J.M., Kevin Park, B., Kordes, C., Kullak-Ublick, G.A., LeCluyse, E.L., Lu, P., Luebke-Wheeler, J., Lutz, A., Maltman, D.J., Matz-Soja, M., McMullen, P., Merfort, I., Messner, S., Meyer, C., Mwinyi, J., Naisbitt, D.J., Nussler, A.K., Olinga, P., Pampaloni, F., Pi, J., Pluta, L., Przyborski, S.A., Ramachandran, A., Rogiers, V., Rowe, C., Schelcher, C., Schmich, K., Schwarz, M., Singh, B., Stelzer, E.H., Stieger, B., Stober, R., Sugiyama, Y., Tetta, C., Thasler, W.E., Vanhaecke, T., Vinken, M., Weiss, T.S.,

406

X. Yang et al. / Toxicology Letters 225 (2014) 401–406

Widera, A., Woods, C.G., Xu, J.J., Yarborough, K.M., Hengstler, J.G., 2013. Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative hepatocyte sources and non-parenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME. Archiv. Toxicol. 87, 1315–1530. Gonzales, P.A., Pisitkun, T., Hoffert, J.D., Tchapyjnikov, D., Star, R.A., Kleta, R., Wang, N.S., Knepper, M.A., 2009. Large-scale proteomics and phosphoproteomics of urinary exosomes. Journal of the American Society of Nephrology: JASN 20, 363–379. Huang, X., Yuan, T., Tschannen, M., Sun, Z., Jacob, H., Du, M., Liang, M., Dittmar, R.L., Liu, Y., Liang, M., Kohli, M., Thibodeau, S.N., Boardman, L., Wang, L., 2013. Characterization of human plasma-derived exosomal RNAs by deep sequencing. BMC Genomics 14, 319. Hussaini, S.H., Farrington, E.A., 2014. Idiosyncratic drug-induced liver injury: an update on the 2007 overview. Expert Opin Drug Saf. 13, 67–81. Kalra, H., Adda, C.G., Liem, M., Ang, C.S., Mechler, A., Simpson, R.J., Hulett, M.D., Mathivanan, S., 2013. Comparative proteomics evaluation of plasma exosome isolation techniques and assessment of the stability of exosomes in normal human blood plasma. Proteomics. 13, 3354–3364. Kucharzewska, P., Christianson, H.C., Welch, J.E., Svensson, K.J., Fredlund, E., Ringner, M., Morgelin, M., Bourseau-Guilmain, E., Bengzon, J., Belting, M., 2013. Exosomes reflect the hypoxic status of glioma cells and mediate hypoxia-dependent activation of vascular cells during tumor development. P. Natl. Acad. Sci. USA 110, 7312–7317. Lasser, C., Alikhani, V.S., Ekstrom, K., Eldh, M., Paredes, P.T., Bossios, A., Sjostrand, M., Gabrielsson, S., Lotvall, J., Valadi, H., 2011a. Human saliva, plasma and breast milk exosomes contain RNA: uptake by macrophages. J. Translational Med. 9, 9. Lasser, C., lsquo, O., Neil, S.E., Ekerljung, L., Ekstrom, K., Sjostrand, M., Lotvall, J., 2011. RNA-containing exosomes in human nasal secretions. American journal of rhinology & allergy 25, 89–93. Levanen, B., Bhakta, N.R., Torregrosa Paredes, P., Barbeau, R., Hiltbrunner, S., Pollack, J.L., Skold, C.M., Svartengren, M., Grunewald, J., Gabrielsson, S., Eklund, A., Larsson, B.M., Woodruff, P.G., Erle, D.J., Wheelock, A.M., 2013. Altered microRNA profiles in bronchoalveolar lavage fluid exosomes in asthmatic patients. J. Allergy Clin. Immunol. 131, 894–903. Llorente, A., Skotland, T., Sylvanne, T., Kauhanen, D., Rog, T., Orlowski, A., Vattulainen, I., Ekroos, K., Sandvig, K., 2013. Molecular lipidomics of exosomes released by PC-3 prostate cancer cells. Biochim. Biophys. Acta 1831, 1302–1309. Masyuk, A.I., Masyuk, T.V., Larusso, N.F., 2013. Exosomes in the pathogenesis, diagnostics and therapeutics of liver diseases. J. Hepatol. 59, 621–625. McGill, M.R., Jaeschke, H., 2013. Apoptosis or necrosis in acetaminophen-induced acute liver failure? New insights from mechanistic biomarkers*. Crit. Care Med. 41, 2653–2654. Mendrick, D.L., 2011. Transcriptional profiling to identify biomarkers of disease and drug response. Pharmacogenomics 12, 235–249. Miyamoto, M., Yanai, M., Ookubo, S., Awasaki, N., Takami, K., Imai, R., 2008. Detection of cell-free, liver-specific mRNAs in peripheral blood from rats with hepatotoxicity: a potential toxicological biomarker for safety evaluation. Toxicol. Sci. : An Official Journal of the Society of Toxicology 106, 538–545. Momen-Heravi, F., Balaj, L., Alian, S., Mantel, P.Y., Halleck, A.E., Trachtenberg, A.J., Soria, C.E., Oquin, S., Bonebreak, C.M., Saracoglu, E., Skog, J., Kuo, W.P., 2013. Current methods for the isolation of extracellular vesicles. Biol. Chem. 394, 1253–1262. Murakami, Y., Toyoda, H., Tanahashi, T., Tanaka, J., Kumada, T., Yoshioka, Y., Kosaka, N., Ochiya, T., Taguchi, Y.H., 2012. Comprehensive miRNA expression analysis in peripheral blood can diagnose liver disease. PloS One 7, e48366. Okubo, S., Miyamoto, M., Takami, K., Kanki, M., Ono, A., Nakatsu, N., Yamada, H., Ohno, Y., Urushidani, T., 2013. Identification of novel liver-specific mRNAs in plasma for biomarkers of drug-induced liver injury and quantitative evaluation in rats treated with various hepatotoxic compounds. Toxicol. Sci.: An Official Journal of the Society of Toxicology 132, 21–31. Poliakov, A., Spilman, M., Dokland, T., Amling, C.L., Mobley, J.A., 2009. Structural heterogeneity and protein composition of exosome-like vesicles (prostasomes) in human semen. The Prostate 69, 159–167.

Raposo, G., Stoorvogel, W., 2013. Extracellular vesicles: exosomes, microvesicles, and friends. J. Cell Biol. 200, 373–383. Regev, A., 2013. How to avoid being surprised by hepatotoxicity at the final stages of drug development and approval. Clin. Liver Dis. 17, 749–767. Ribeiro, R.M., Layden-Almer, J., Powers, K.A., Layden, T.J., Perelson, A.S., 2003. Dynamics of alanine aminotransferase during hepatitis C virus treatment. Hepatology 38, 509–517. Royo, F., Falcon-Perez, J.M., 2012. Liver extracellular vesicles in health and disease. J Extracell Vesicles. 1. Royo, F., Schlangen, K., Palomo, L., Gonzalez, E., Conde-Vancells, J., Berisa, A., Aransay, A.M., Falcon-Perez, J.M., 2013. Transcriptome of extracellular vesicles released by hepatocytes. PloS One 8, e68693. Shi, Q., Hong, H., Senior, J., Tong, W., 2010. Biomarkers for drug-induced liver injury. Expert Rev. Gastroenter. Hepatol. 4, 225–234. Shi, Q., Yang, X., Greenhaw, J., Salminen, W.F., 2011. Hepatic cytochrome P450s attenuate the cytotoxicity induced by leflunomide and its active metabolite A77 1726 in primary cultured rat hepatocytes. Toxicol. Sci.: An Official Journal of the Society of Toxicology 122, 579–586. Shi, Q., Yang, X., Mendrick, D.L., 2013. Hopes and challenges in using miRNAs as translational biomarkers for drug-induced liver injury. Biomarkers Med. 7, 307–315. Starkey Lewis, P.J., Dear, J., Platt, V., Simpson, K.J., Craig, D.G., Antoine, D.J., French, N.S., Dhaun, N., Webb, D.J., Costello, E.M., Neoptolemos, J.P., Moggs, J., Goldring, C.E., Park, B.K., 2011. Circulating microRNAs as potential markers of human druginduced liver injury. Hepatology 54, 1767–1776. Street, J.M., Barran, P.E., Mackay, C.L., Weidt, S., Balmforth, C., Walsh, T.S., Chalmers, R.T., Webb, D.J., Dear, J.W., 2012. Identification and proteomic profiling of exosomes in human cerebrospinal fluid. J. Translational Med. 10, 5. Tolosa, L., Pinto, S., Donato, M.T., Lahoz, A., Castell, J.V., O‘Connor, J.E., Gomez-Lechon, M.J., 2012. Development of a multiparametric cell-based protocol to screen and classify the hepatotoxicity potential of drugs. Toxicol. Sci.: An Official Journal of the Society of Toxicology 127, 187–198. Turchinovich, A., Weiz, L., Langheinz, A., Burwinkel, B., 2011. Characterization of extracellular circulating microRNA. Nucleic Acids Res. 39, 7223–7233. van der Pol, E., Boing, A.N., Harrison, P., Sturk, A., Nieuwland, R., 2012. Classification, functions, and clinical relevance of extracellular vesicles. Pharmacol. Rev. 64, 676–705. Wang, K., Zhang, S., Marzolf, B., Troisch, P., Brightman, A., Hu, Z., Hood, L.E., Galas, D.J., 2009. Circulating microRNAs, potential biomarkers for drug-induced liver injury. P. Natl. Acad. Sci. USA 106, 4402–4407. Welker, M.W., Reichert, D., Susser, S., Sarrazin, C., Martinez, Y., Herrmann, E., Zeuzem, S., Piiper, A., Kronenberger, B., 2012. Soluble serum CD81 is elevated in patients with chronic hepatitis C and correlates with alanine aminotransferase serum activity. PloS One 7, e30796. Wetmore, B.A., Brees, D.J., Singh, R., Watkins, P.B., Andersen, M.E., Loy, J., Thomas, R.S., 2010. Quantitative analyses and transcriptomic profiling of circulating messenger RNAs as biomarkers of rat liver injury. Hepatology 51, 2127–2139. Witwer, K.W., Buzas, E.I., Bemis, L.T., Bora, A., Lasser, C., Lotvall, J., Nolte-, rsquo, t Hoen, E.N., Piper, M.G., Sivaraman, S., Skog, J., Thery, C., Wauben, M.H., Hochberg, F., 2013. Standardization of sample collection, isolation and analysis methods in extracellular vesicle research,. J Extracell Vesicles 2. Xu, J.J., Henstock, P.V., Dunn, M.C., Smith, A.R., Chabot, J.R., de Graaf, D., 2008. Cellular imaging predictions of clinical drug-induced liver injury. Toxicol. Sci.: An Official Journal of the Society of Toxicology 105, 97–105. Yamaura, Y., Nakajima, M., Takagi, S., Fukami, T., Tsuneyama, K., Yokoi, T., 2012. Plasma microRNA profiles in rat models of hepatocellular injury, cholestasis, and steatosis. PloS One 7, e30250. Yang, X., Greenhaw, J., Shi, Q., Su, Z., Qian, F., Davis, K., Mendrick, D.L., Salminen, W.F., 2012. Identification of urinary microRNA profiles in rats that may diagnose hepatotoxicity. Toxicol. Sci.: An Official Journal of the Society of Toxicology 125, 335–344. Yun, J.W., Lee, T.R., Kim, C.W., Park, Y.H., Chung, J.H., Lee, Y.S., Kang, K.S., Lim, K.M., 2010. Predose blood gene expression profiles might identify the individuals susceptible to carbon tetrachloride-induced hepatotoxicity. Toxicol. Sci.: An Official Journal of the Society of Toxicology 115, 12–21.